SShogo Bench

GPT-4o (2024-08-06)

GPT model for general reasoning, writing, coding, and tool-assisted tasks

Shogo Agent Index

Supported#90 of 92
12.2range 10–14.5 · 34% of weight measured
Agentic coding
10
34% measured
Long-horizon autonomy
22
50% measured
Knowledge-work agents
12
28% measured
Computer use
2
estimated
Security & context learning
11
50% measured

Price

Input
$2.27/M
Output
$9.09/M
Cache read
$1.14/M
3:1 blend
$3.98/M
Agent-effective
$1.68/M

Cheapest host for agents: Azure at $1.85/M.

Agent-effective = 20 input tokens per output token, 85% cache hits. Why.

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verified31.0%8%source
Terminal-Bench 2.0est.8%predicted 2
DeepSWEest.6%predicted 4
FrontierSWEest.5%predicted 0
CursorBenchest.4%predicted 12
GSO-Bench0.0%4%OpenHands
METR Time Horizons40.8%10%source
Vending-Bench 2est.6%predicted 0
PostTrainBenchest.4%predicted 8
APEX-Agentsest.8%predicted 5
GDPval9.9%7%source
Remote Labor Indexest.5%predicted 0
DeepResearch Benchest.5%predicted 38
OSWorld 2.0est.6%predicted 0
OSWorldest.4%predicted 6
Cybench12.5%5%source
CL-benchest.5%predicted 10

Other benchmarks

BenchmarkCategoryScoreSource
Aider PolyglotCoding45.3%source
ARC-AGIReasoning4.5%source
ARC-AGI-2Reasoning0.0%source
BALROGLong-horizon32.3%source
Chess PuzzlesReasoning9.0%source
CritPtReasoning0.0%source
DTBenchKnowledge64.5%source
Fiction.LiveBenchLong context66.7%source
FrontierMath Tiers 1-3Math0.4%source
GeoBenchMultimodal71.0%source
GPQA DiamondReasoning49.2%source
Humanity's Last ExamReasoning2.7%source
LMCAKnowledge16.6%source
MATH Level 5Math53.3%source
MMLULegacy88.1%source
OTIS Mock AIMEMath6.4%source
SimpleBenchReasoning17.8%source
SimpleQA VerifiedKnowledge26.0%source
TheAgentCompanyKnowledge work8.6%source
VPCTMultimodal40.0%source

API hosts

Live prices and uptime for every host serving GPT-4o (2024-08-06) via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
Azure$2.50$10.00$1.25$1.85100.0%—
OpenAI$2.50$10.00$1.25$1.8599.9%—